We propose a harmacoepidemiologic study of 5-alpha-reductase inhibitors (5ARIs) to address emerging safety concerns (permanent sexual side effects, depression) with use of these marketed drugs, presently being considered for expanded use in healthy men as agents of prostate cancer chemoprevention. In order to cost- efficiently deliver needed safety information, this research will examine the occurrence of adverse events (AEs) via analysis of a large electronic medical records database, the UK-based Clinical Practice Research Datalink (CPRD). The CPRD is the most widely-used resource for pharmacoepidemiologic research in the world. In preliminary analyses of this very large database (11M patients, 45M patient-years), we have identified over 58,000 users of the drug class in question with on average approximately 10 years of follow-up time. In addition to studying risk of events, the proposed study will be the first to study symptom persistence in relation to both types of drugs in the class (finasteride and dutasteride), and will be the largest epidemiologic study of its kind to date. We propose to examine the occurrence of 4 thematic groups of AEs: 1) erectile dysfunction (ED); 2) non- ED sexual dysfunction (ejaculatory and psychosexual dysfunction, low libido, Peyronie's disease); 3) breast outcomes (benign/malignant breast diseases), and 4) depression, using longitudinal epidemiologic approaches. In addition to a retrospective cohort study focused on incident events, we will examine the persistence/permanency of sexual dysfunction outcomes and depression following drug cessation, and risk factors for permanency. Our analyses are designed such that men under various drug treatments for benign prostatic hyperplasia (BPH) are compared (finasteride, dutasteride, and alpha-blockers) to account for the underlying influence of BPH on outcomes. We will examine potential confounding in our analysis using the wide variety of medical and behavioral factors available in CPRD, including use of novel methods such as high-dimensional propensity score analysis. Our research team is comprised of experts in the epidemiology of male aging, pharmacoepidemiology relating to sexual function and lower urinary tract symptoms, and clinical treatment of BPH. The PI has immediate access and >25 years of experience in pharmacoepidemiologic utilization of the CPRD. Other strengths of our proposed study include its large size, use of medical records (avoiding the need to rely on patient recall o prescriptions and avoidance of problems with use of medical claims), longitudinal design, the ability to compare 5ARIs to other treatments, and the richness of the underlying dataset for confounder control. The public health contribution of the study is to quickly to inform the safety profile of the 5ARI class for currently-approved indications before further expansion to healthy men for chemoprevention.